inspirehep.net vs TensorFeed

Same instrument, two spec sheets — measured, not claimed.

Uptime · 30d
Uptime · 90d100%100%
Uptime · 30d100%100%
P50 · ms514306
P95 · ms923752
Authnonenone
CORSyesyes
HTTPSyesyes
Card requirednono
Commercial useunclearunclear
Data licenseUnverifiedUnverified
Free tierFree — limits not publishedFree — no key (static JSON feed endpoints)
Rate limitUnpublished120 req/window · 119 remaining · resets 60
In directory since2026-07-052026-07-05
operationalpartialdownno data

inspirehep.net vs TensorFeed: common questions

Which is more reliable, inspirehep.net or TensorFeed?

Both are neck-and-neck — inspirehep.net and TensorFeed each measure 100% uptime over 90 days on our probe schedule. Reliability here is verified from our own scheduled checks; use the 30-day bars above to see which has been steadier lately.

Which is faster, inspirehep.net or TensorFeed?

TensorFeed has the lower median latency in our checks — inspirehep.net responds in 514 ms versus TensorFeed at 306 ms (P50). Tail latency (P95) is in the table above; for most workloads the median is the number that shapes how the API feels.

Do inspirehep.net and TensorFeed need an API key?

Neither needs a paid key — inspirehep.net is callable with no signup, and TensorFeed is callable with no signup. Both are quick to prototype with; rate limits still apply.

Can I call inspirehep.net and TensorFeed from the browser?

Yes — both inspirehep.net and TensorFeed send CORS headers over HTTPS, so front-end code can fetch either directly with no backend proxy. That makes them easy to swap in a client-side app while you compare responses.

Are inspirehep.net and TensorFeed free for commercial use?

inspirehep.net has unclear commercial terms, and TensorFeed has unclear commercial terms. We track service terms and the data license as separate fields — see the Commercial use and Data license rows above, and confirm both before shipping either in a paid product.